On Thu, Nov 20, 2014 at 6:35 AM, White, Jason J <jjwhite@ets.org> wrote:
> This discussion has been informative, but it motivates a call for
> clarification of the semantic differences between the proposed aria-current
> state, aria-selected and aria-activedescendant. As I understand the
> differences:
>
> aria-activedescendant implies that the element has the keyboard focus.
>
No, this isn't necessarily true at all. Suppose you have a listbox with 10
choices. The 5th one is highlighted. If the listbox itself is focused, then
the active descendant tells you that the 5th item is the one that's active
and is the focused item within the focused control. However, if focus
leaves the listbox, the 5th item may still be highlighted; it's still the
active item within the list and it's where focus would go if you focused
the listbox again.
For that reason, I'm leaning towards calling this proposed new attribute
aria-active. It indicates that an item within a group is the active one,
whether or not it currently has focus or not. It tells you something about
what happens when the control does have focus.
In the case of the listbox, if it's a single-select listbox then perhaps
"active" and "selected" mean the same thing. But if it's a multi-select
listbox, only one can be active but many can be selected.